Hi: It is not good graphics for a legend to overlap data points, so let's see how one can get said legend away from the data.
One approach in base graphics is to extend the y-axis limits and to place the legend in the open area. It would probably be better to split the legend into two halves and place separate legends on both ends, but I'll let you do that as R homework. The following is a small modification of David's yeoman work; the dotted line isn't really necessary, but it provides a visual separation between the plot and the legend. (And, to echo David's remarks about the code blob, I hope this is easier to read and understand...) x<-runif(100) y<-runif(100) color<-c(rep(1,15),rep(2,20),rep(3,15),rep(4,10),rep(5,10),rep(6,15),rep(7,10),rep(8,5)) # Combine vectors into a data frame and remove original variables from the workspace mydf <- data.frame(x = x, y = y, color = color) rm(x, y, color) # Legend elements mylegend <- expression(scriptstyle("All 3 outliers"), scriptstyle("Presence of 2 outliers (6,263 and 6,720)"), scriptstyle("Presence of 2 outliers (3,471 and 6,720)"), scriptstyle("Presence of 2 outliers (3,471 and 6,263)"), scriptstyle("Presence of outlier 6,720"), scriptstyle("Presence of outlier 6,263"), scriptstyle("Presence of outlier 3,471"), scriptstyle("No outlier") ) mypch <- c(22,3,3,3,19,19,19,23) mycol <- c("black","blue","green","red","red","green","blue","black") plot(y ~ x, data = mydf, col = color, pch = mypch[color], ylim = c(0, 1.5), xlab='Sample Mean', ylab='Bootstrap Variance Estimators', main = '') legend('topleft', leg = mylegend, col = mycol, pch = mypch, lty = c(1, 1), xjust=0, yjust=0, bty="n") abline(1.02, 0, lty = 'dotted') Since I have to contribute something novel, here's a lattice version that David suggested would make a good alternative, and I believe it does, but that's my opinion - judge it for yourself. library(lattice) # Reconstitute the original legend text items mylegend2 <- c("All 3 outliers", "Presence of 2 outliers (6,263 and 6,720) ", "Presence of 2 outliers (3,471 and 6,720) ", "Presence of 2 outliers (3,471 and 6,263) ", "Presence of outlier 6,720", "Presence of outlier 6,263", "Presence of outlier 3,471", "No outlier" ) # Generate code for a legend key mykey <- list(space = 'top', columns = 2, text = list(mylegend, cex = 0.8), points = list(pch = mypch, col = mycol, cex = 0.8)) xyplot(y ~ x, data = mydf, col = mydf$color, pch = mypch[mydf$color], xlab='Sample Mean', ylab='Bootstrap Variance Estimators', main = '', key = mykey) HTH, Dennis On Tue, Sep 21, 2010 at 10:28 AM, Mestat <mes...@pop.com.br> wrote: > > Hi Denis, > Check it out my code... This is not my real data... > I would like to manage the size of the legend... Set the legend smaller > than > it is, because on my real data, the legend is over the values... > Thanks for the help... > > x<-runif(100) > y<-runif(100) > > color<-c(rep(1,15),rep(2,20),rep(3,15),rep(4,10),rep(5,10),rep(6,15),rep(7,10),rep(8,5)) > > > plot(x,y,col=c("black","blue","green","red","red","green","blue","black")[color],pch=c(22,3,3,3,19,19,19,23)[color],main=' > ',xlab='Sample Mean',ylab='Bootstrap Variance Estimators') > legend("topleft", legend=c("All 3 outliers","Presence of 2 outliers (6,263 > and 6,720)","Presence of 2 outliers (3,471 and 6,720)","Presence of 2 > outliers (3,471 and 6,263)","Presence of outlier 6,720","Presence of > outlier > 6,263","Presence of outlier 3,471","No outlier"), col = > c("black","blue","green","red","red","green","blue","black"), pch = > c(23,19,19,19,3,3,3,22), lty=c(1,1), xjust=0, yjust=0, bty="n") > > -- > View this message in context: > http://r.789695.n4.nabble.com/Size-of-the-legend-tp2547601p2549063.html > Sent from the R help mailing list archive at Nabble.com. > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.